Net gains: How big data is helping tennis fans and media get to the point

IBM has been working at Wimbledon for 33 years with an increasingly complex operation that analyses every game and player

Novak Djokovic returns the ball to Nick Kyrgios during their men's singles final at Wimbledon on Sunday July 10th. Photograph: Sebastien Bozon/AFP via Getty Images
Novak Djokovic returns the ball to Nick Kyrgios during their men's singles final at Wimbledon on Sunday July 10th. Photograph: Sebastien Bozon/AFP via Getty Images

The trophies have been lifted, the Pimm’s drunk and the dust has just about settled on Centre Court at the All England Lawn Tennis & Croquet Club in Wimbledon.

Serbian champion Novak Djokovic lifted the trophy for the seventh time on Sunday, beating Australian Nick Kyrgios to the title, while the day before saw Kazakhstani player Elena Rybakina snatch victory over Tunisian Ons Jabeur in the women’s final.

The 2022 tournament was the first full-capacity event since 2019, with Covid-19 cutting attendances in 2021 and the tournament cancelled entirely the year before. But while all eyes were on the action on court, there was a lot more going on in the background.

In the basement of the broadcast centre, tucked away behind Court 18, lies the nerve centre for a sophisticated operation that collects and analyses data from every game.

READ MORE

That means monitoring every shot, rally and fault from different angles, the type of shot, success rate and so on. This huge pool of data underpins the insights and statistics that are fed to the Wimbledon website and app, and supply data to the broadcasters covering the tournament.

This is where tech giant IBM carries out its operations. The company has worked with the All England Lawn Tennis Club on the championship for three decades, with 2022 marking the 33rd year of the technology partnership. Since 1990, IBM has provided technology and consulting services to bring innovation to what is viewed as a traditional organisation, using the technology IBM has developed for its business customers to help Wimbledon on its quest for digitalisation, powering its fan platforms and keeping the tournament’s systems secure.

And evidence of its success is all around. In the past few years, the tournament has seen the introduction of AI-powered predictions for matches, the power index that ranks players, recommendations and AI-powered highlight reels that are generated automatically.

It’s all in the name of fan engagement, attracting more people to the tournament through its digital platforms.

This network of rooms in the media centre, out of the heatwave that hit Wimbledon, is where the operations are co-ordinated. Banks of screens display stats and matches on all the courts being played at Wimbledon. Shots are examined. On the wall of one room is a commemoration of the fastest serve recorded by IBM at Wimbledon: 148 miles per hour, served by Taylor Dent in 2010; second to that was Andy Roddick.

If statisticians watching the games are unsure about a particular shot, they may call down here for clarification from the people behind the screens, who can replay all the points on all the courts around Wimbledon.

Over the course of the championship, IBM collects and analyses 125,000 data points, including about 2,000 per match. How the ball bounces, for example, or a player’s gestures during the match. In total, it has collected more than nine million data points at Wimbledon, an astonishing amount of data that IBM has at its fingertips.

“That data is what drives the statistics that go on the screens when you watch the BBC or coverage,” says Simon Boyden, digital architect with IBM. “All of that data is coming from this kind of nerve centre.”

In 2017, IBM Watson was brought into the mix. That is IBM’s artificial intelligence for business, which helps organisations predict future outcomes and automate complex processes. Now, it can also help make predictions for matches.

“Every player ends up with a whole set of structured data about them,” said Boyden. “We have a historical archive of it. What we’ve been really focusing on over the last few years is using Watson to tap into some of the unstructured data.”

A combination of IBM Watson Discovery and IBM Cloud analyses performance data, looks through media commentary, and measures player momentum.

Last year saw the introduction of the IBM Power Index, which uses natural language processing to analyse player performance before and during the championship. The same system was used for prematch insights, a preview sheet generated by AI that compares players head to head.

This year’s big innovation is the ability in the app for fans to have their say on how they think the matches will go, and compare their thoughts with those of other fans and the AI-generated predictions supplied by IBM.

For fans, these match predictions and the ability to have their say has been a firm favourite.

Although Watson hits the mark more often than not in its predictions, it is not infallible. A quick scroll through the app shows its predictions and how close to the mark they were. Although it correctly predicted the result of the men’s and women’s semi-finals that were played, the finals were a different story.

Before the players even set foot on the pitch, Watson had predicted a win for Kyrgios — the wrong result, in the end, and proof that technology isn’t infallible. The IBM system had taken a number of things into account, including Kyrgios’s previous record against Djokovic; he had won both the previous two outings against the champion.

In contrast, the fan predictions swung the other way, proclaiming Djokovic would be the winner.

For the women’s final the previous day, the system had predicted a win for Jabeur, but it was Rybakina who lifted the trophy.

That’s half the fun of it, though.

“Ultimately, it’s about provoking debate,” explains Boyden. “One of the things that we’ve added for this year is this idea of explainable AI. The challenge with AI is often that you have a very complicated big black box that data goes into and results come out. One of the trends we’re seeing is around explainable AI — why did the model think this?”

Artificial intelligence is also taking over some of the coverage. It can automatically create highlight reels for critical match events as quickly as possible. It uses three main factors to decide if an event is worth highlighting: crowd noise, player gestures, and the match situation, such as if it is a game or match point. A process that would have been done manually a few years ago is now being controlled by algorithms.

“Leveraging technology to help fans become more informed, engaged and involved throughout the Wimbledon fortnight is at the core of our strategy to ensure we are leveraging innovation to keep Wimbledon relevant and deliver outstanding digital experiences for fans, wherever they may be,” said Alexandra Willis, communications and marketing director at the All England Club. “Core to these experiences is our ambition to help fans get closer to Wimbledon by understanding which players to follow and analyse, and inviting them to get involved with new match predictions and insights features, alongside our extensive scoring, news and video content across our channels.”

The whole operation is run on IBM Cloud and hybrid cloud technologies, with more and more functions being shifted to the cloud. Although some functions are still done on site for operational reasons, recent years have seen the balance shift to the cloud. The system also has to be robust enough to take a surge in traffic without collapsing, and secure enough to keep out unwanted intruders.

Keeping innovation alive at the tournament is a commitment. And although the tournament has just wrapped up, the two organisations are already working on innovations for 2023.

Twice a year, in spring and autumn, IBM’s digital and design consultancy arm IBM iX holds workshops with IBM and All England Lawn Tennis & Croquet Club representatives to discuss where they will go next. Spring is when the new ideas are put forward and discussed; autumn is when they decide what will be implemented the following year.

While the focus may be on driving innovation, though, doing things right is just as much of a priority, says Boyden.

“It’s quality. You’ve got to get this right.”